Re: [Sedate] Distinguishing "optional extra" vs "you need to understand this to process the date accurately" extensions

[John C Klensin <[e-mail]>]

--On Thursday, December 9, 2021 15:59 +0100 Carsten Bormann
<[e-mail]> wrote:

> On 2021-12-09, at 14:04, Bron Gondwana
> <[e-mail]> wrote:
>> 
>> So I wonder if we need a syntax to say "if this extension is
>> not recognized by the receiving system, it SHOULD throw an
>> error and MUST NOT expect the date to be exact" - either a
>> reserved namespace for such extensions, or some other
>> syntactical sugar.
> 
> Protocols typically have a way to distinguish "critical"
> (or "must-understand") options from "elective"
> options.  Sometimes, the distinction is baked into the option
> code, sometimes there is separate place (which also allows the
> same option to be used in both ways).  Our syntax should
> provide an easy way to do the latter, as in:
> 
>    1996-12-19T16:39:57-08:00[America/Los_Angeles][u-ca=hebrew]
> (Elective, ignore option if not understood)
> 
>    1996-12-19T16:39:57-08:00[America/Los_Angeles][u-ca!hebrew]
> (Critical, must understand/must reject date if not understood)
> 
> Well, we need better examples...

Carsten,

Using your examples and adapting your syntax (purely as
examples), would you also expect/allow

    1996-12-19T16:39:57-08:00[!America/Los_Angeles][u-ca=hebrew]
How about 

2022-12-19T16:39:57!-08:00[!America/Los_Angeles][!u-ca=hebrew]
which might mean "that time, somewhere"?

:-(
Standing one the edge and looking down the slippery slope here.

best,
   john